Guiding due to MOUNT deflection(mechanical flexure) (ie periodic error) is more primary a concern than OTA flex at sidereal rate. With PEC dialed out next comes perceived change in image from atmosphere, and while .2-.25 would be ideal, atmospheric conditions can make that goal unobtainable in many places. Least would be concern about flex due to OTA design.
In my experience flexure of the mount causing image shift is far more serious than believing the OTA is somehow sagging. More often the discussion is external guide scope delta change vs OTA delta change. Using an OAG or OOAG would mitigate that problem IF the individual felt their system had enough backfocus adjustment to make use of such systems and achieve their goals. Yet an external guide system is more than adequate if proper attention is paid to selection of components and considerations of where the delta change might occur in the system. Those of us who guided by eye in the film days can attest to how much better images are today with guide cameras used both in OAG and Guidescope configurations. The key is understanding the FOV equivalence between guide and main scope and any atmospherics. Todays more sensitive guide cameras can pick up the delta change much quicker making even a piggyback guider capable of handling big SCTs. In fact it becomes necessary if the target FOV through the OTA contains no reasonable guide stars to rely on a wider angle guider.
That sure hasn't been my experience with the 4x C14s that I've imaged with--on two different mounts. My current mount is an AP1600 with encoders and the primary factor that limits pointing accuracy is mechanical flexure due to the OTA and its internal components. Here's how I know that: You can easily assess gross flexure by doing a polar field rotation. Mount a small, solid refractor on your mount, point at the pole and run the mount to the east side of the pier. Then open the shutter and run the RA to the west side of the pier. If the trails are not circular, there's flexure in your mount; otherwise it's good. Now mount your SCT and do the same thing. If your mount is good, you'll see on a large scale how much flexure is directly related to the OTA. I've attached a PFR for my C14 system, which is on a good mount that's rated for about 3x the weight that it actually carries. Look carefully at the trailed stars and you'll see that the system is clearly flexing as it moves through the RA angle. This turns out to be the limiting factor for pointing errors even when atmospheric correction is applied to the encoder controlled position. In general, this scope is only good to about 0.3 degrees over the whole sky--and that's all due to mechanical flexure. This scope is pretty bad and some might be better; but, none are perfect--particularly when you look at any scale less than an arc-second.
An external guide scope will indeed work with these scopes, but you'll never achieve very good results with any exposure longer than a minute (or maybe two)--and even then, the performance will be hit and miss. In order to achieve FWHM stars below about 5", you have to guide through the telescope itself--and that's always due to mechanical flexure between the two scopes. Keep in mind that mechanical flexure of only 0.25" across the length of a C14 is only about 34 millionths of an inch (0.9 microns or slightly less than 2 wavelengths of light.) You can't hardly touch that tube without deflecting it by that much! With my ONAG system, I regularly achieve 0.25" guiding with FWHM below 2" and I've seen as good as 1" when the conditions permit using 20 minute exposures. That level of performance is impossible using a guide scope.
PS I too started guiding film by eye in the late 60's and it was terrible by today's standard.